Tumor hypoxia has been shown to decrease tumor control and survival in patients with head and neck squamous cell carcinoma (HNSCC) treated with radiotherapy. With the development of therapies targeted against hypoxia such as Tirapazamine (TPZ) and HIF-1oc inhibitors, there is a need for a simple and non- invasive way to identify patients who would benefit from such treatments. Presently, the most widely accepted method for detecting tumor hypoxia is the Eppendorf pO;. histograph, which is an invasive technique that requires a high level of patient cooperation. In addition, it cannot be used for deeply seated tumors. We have recently shown that Osteopontin (OPN), a secreted protein, whose expression can be measured in patient blood via a simple ELISA test, can be used as a surrogate marker for tumor hypoxia in a small group of HNSCC patients treated with chemoradiotherapy. Plasma OPN level correlated not only with tumor pO2 measurements but also with treatment outcomes (freedom from relapse and overall survival). The HeadSTART trial, which is an international phase III randomized trial evaluating the effectiveness of adding TPZ, a hypoxic cytotoxin, to conventional chemoradiotherapy in patients with advanced HNSCC provides a unique opportunity to validate the role of OPN as a secreted hypoxia marker. In this study, 850 HNSCC patients are randomized to concomitant radiation, cisplatin and TPZ versus concomitant radiation and cisplatin alone. Pretreatment plasma samples of these patients will be analyzed for OPN levels using an optimized ELISA system and compared to treatment outcomes in terms of failure-free survival and overall survival. If OPN is indeed a secreted marker for tumor hypoxia, the addition of TPZ would be predicted to benefit only those patients with elevated pretreatment OPN. Similarly, the elimination of hypoxic tumor cells by TPZ will diminish the prognostic impact of OPN in the experimental arm when compared to the control (non-TPZ) arm. We plan to correlate the expression of OPN with tumor expression of 4 known tissue makers for hypoxia (HIF-1cc, VEGF, CA IX and Glut-1) and 2 novel markers (Galectin-1 and Ephrin A1). We will also correlate plasma OPN levels with tumor hypoxia that is determined from pretreatment [18]Fluoroazomicin- arabinofuranoside (18FAZA) PET scans in a subset of patients. Correlation of OPN to hypoxia-specific PET scans and tissue markers will provide 2 different approaches to validate OPN as secreted hypoxia markers in HNSCC. As a secondary objective, we will study the expression and prognostic impact of 3 other potential secreted markers for hypoxia, which are plasma VEGF, IL-8 and CTGF. In addition, we will explore the use of MALDI-ToF proteomic and several sophisticated statistical softwares to identify a global proteomic signature for predicting relapse in patients treated with and without TPZ. We will also conduct proteomic analysis of serial plasma samples obtained from the same patients before and after TPZ to identify novel hypoxia-related secreted proteins that can server as future biomarkers and targets for new therapy.